When several containers serve one app, each can report its own metrics. PromQL is Prometheus’s query language for turning those time series into useful answers: select a metric, filter by labels, and aggregate values across containers. section 9.3
Count requests across containers
Run `sum(access_log_total) without(hostname, instance)` in the Prometheus UI. The sum combines request counts, while `without` removes the labels that distinguish individual containers. Add a graph to see how the total changes over time as you send requests. section 9.3
sum(image_gallery_requests_total{code="500"}) without(instance)
This dashboard query first selects requests with response code 500, then sums them without separating results by instance. It provides a focused view of HTTP errors across the app. section 9.4
Use the Prometheus UI to check scrape targets and work out queries, then put those queries into Grafana panels. Filtering and aggregation are enough for useful dashboards; functions for rates, changes over time, and subqueries can wait. section 9.3 section 9.4
